Using Simulation Techniques within the Design of a Reconfigurable Architecture for Fail-Operational Real-Time Automotive Embedded Systems

In the automotive industry one clear goal is to reach the fourth SAE level of driving automation. Instead of having a redundant E/E-architecture we propose a dynamically reconfigurable one. With this paper we introduce two new simulation techniques on how to extract two important values for reconfiguration. The approach is based on an example for the loss of steering control functionality. The first value is the overall system reconfiguration time that is recognized by the human driver in case of a loss of steering control functionality. The second value is the maximum reconfiguration time possible to still perform a given maneuver. The results of both values are compared to a previous study. This leads to end-to-end reconfiguration times for future development of fail-operational self-reconfigurable real-time automotive embedded systems. A method for simulating loss of functionality and deriving specification relevant data is given to assist the automotive industry reaching fail-operational systems. It will also help in introducing dynamical reconfiguration for a fail-operational automotive E/E-architecture as the next step.